A flexible approach for normal approximation of geometric and topological statistics (Q6589583)

The authors derive normal approximation results for a class of stabilizing functionals of binomial or Poisson point process, that are not necessarily expressible as sums of certain score functions. So, they give an affirmative answer to the following question: сan one obtain presumably optimal bounds for general functionals that automatically result in presumably optimal bounds when specialized to the case of functionals that can be expressed as sums of score functions? The term ``presumably optimal'' is used to refer to the case when the order of the normal approximation is the same as that of a sum of i.i.d. random variables. The new approach is based on the idea of using the add-one cost operator at two scales. In contrast to previous works, the authors directly simplify the evaluation of the iterated add-one cost operators. In the main theorems they provide normal approximation results for functionals of Poisson and binomial point processes, respectively. Then the applicability of their approach is illustrated by deriving normal approximation results for several geometric and topological statistics. Specifically, normal approximation results are obtained for the total edge length of \(k\)-nearest neighbor graphs and weighted \(k\)-nearest neighbor graph appearing as Shannon entropy estimators. Further, the Euler characteristic is considered, which is an elementary statistics widely used in the field of topological data analysis. Finally, the applicability of the author's approach is discussed in connection to the minimal spanning tree problem, by recovering existing results.